Ke-young Chu and Andrew Feltenstein
Modern countries have all, almost without exception, suffered periods of inflation. In general these inflations have been controllable, both in their severity and in their duration, but there have been certain instances in which the acceleration in the rate of increase in prices has been seemingly unmanageable. Perhaps the most famous such case, and one of the few instances of true hyperinflation, was the German inflation of 1923, during which the rate of increase in prices rose to a peak of more than 30,000 per cent a month and the economy was reduced to functioning virtually as a barter economy.
In the past 20 years there have been no examples of such violent inflation, but there have been a few countries where for a significant period of time the rate of price increase seemed to be genuinely out of control. Three of the most recent examples were Chile in 1973, when inflation reached an annual rate of more than 650 per cent; Indonesia in 1966–67, with rates of increase in prices of over 170 per cent annually; and Argentina in 1973–76, when the rate of inflation had risen to 50 per cent a month by the spring of 1976 and to almost 350 per cent during the course of that year. The case of Argentina is particularly interesting because a number of the country’s economic policies had a significantly adverse impact upon inflation although, paradoxically, they were often designed precisely to reduce the rate of increase in prices. This article summarizes the results of an analysis of the relationship between inflation, price controls, the money supply, and real wages in Argentina for the period 1963–76.
Price controls were a major characteristic of Argentine economic programs implemented between 1963 and 1976—aimed at both controlling inflation and real wages. In 1963 the economy was operating without any significant controls at a rate of inflation that was approximately the postwar average of about 24 per cent a year; between 1964 and 1966, however, the Government, dissatisfied with this rate of price increase, initiated a system of selective direct controls on the prices of the public sector and of certain private industries, such as agriculture. This program did keep prices down initially, but the rate of inflation eventually rose again, and the system was abandoned at the end of 1966. A new Government introduced an anti-inflation program that simultaneously applied restraints to both wages and prices. In support of these measures, the Central Government significantly reduced its deficit, which had been a major source of monetary expansion and hence of inflationary pressures. During 1967 and 1968 certain selective price controls, which attempted to stem the rate of increase in costs, continued to be used. Between 1968 and 1969 Argentina experienced the lowest rates of inflation in recent history while continuing to reduce its reliance on price controls (see Chart 1).
In 1971 this period of relatively low inflation came to an end. Fiscal policy became expansionary and wage programs were relaxed; in addition, in an unsuccessful attempt to slow the rise in prices, firms were not allowed to pass on increases in wage costs to their output prices. In 1973, the Peron Government returned to power in Argentina, bringing with it the Social Contract, a program whose primary goal was to raise the general level of real wages and to redistribute income. Price controls were a key component of this program: in June 1973, prices for a large number of consumer goods were lowered by 7 to 20 per cent, and, at the same time, nominal wages were raised by an average of 20 per cent. As a result, business profits were severely squeezed, shortages and black markets developed, and by April 1974 there was sufficient pressure to compel the Government to relax its controls. Since wages continued to rise, an inflationary spiral began. This weakened the Government’s fiscal position, as tax receipts failed to keep up with inflation, and there were price distortions in those sectors that were still subject to controls. By March 1976 the rate of inflation had reached 50 per cent a month. After a new Government removed all price controls, however, inflation decelerated sharply and prices rose by only 63 per cent in the second half of the year, compared to a rise of 175 per cent in the first half. At the same time, however, real wages declined sharply.
Causes of inflation
Two studies were made of inflation in Argentina for the period between 1963 and 1976 to test a series of hypothetical links between price controls, government deficits, inflation, and real wages. (The models upon which the analysis is based are described in the box.) This analysis shows that there were essentially two main causes of the inflation—both of which had an inflationary impact because domestic credit, and hence the money supply, were automatically linked to demand. The first cause was the budget deficit of the Central Government. The size of the deficit was usually set in real terms, as a percentage of gross domestic product (GDP) for example. As the general price level in the country increased, the nominal level of the deficit had to be raised, and this led to an inflated deficit that was generally financed on the basis of a passive expansion in Central Bank credit. At the same time, taxpayers were often allowed to delay paying fixed amount tax obligations, and, again because of inflation, these delays significantly reduced government revenues in real terms, leading ultimately to a further increase in the deficit.
The second, and less apparent, cause of inflation during this period was the very instrument that the Government often used to attempt to check the rise in prices: namely, price controls. The Government frequently turned to selective price controls not only to keep down the prices of key commodities but also in an attempt to control inflation and maintain high real wages. At the same time, the Government subsidized those public and private firms that had made losses because of price controls, either by direct transfer payments or by low interest rate loans made through the banking system. As a result, distortions in sectoral prices and wage-price relationships in key industries gave rise to an excessive creation of credit by the banking system. This ultimately led to a situation in which it was extremely difficult for the Government to defend its original price targets for any length of time.
Chart 1Quarterly rate of inflation and rate of change in average nominal wage rate in industrial sector
The analysis estimated the distortions in the price structure of the major sectors in the Argentine economy between 1963 and 1976. It was found that, with a few notable exceptions, these sectors had had their output prices held down considerably below the levels at which they could show a profit during 1973–75, which was the period of the Social Contract. Agriculture, for example, experienced considerable losses during those years, because controls kept the selling prices of agricultural goods below production costs. By the end of 1976, however, when price controls had been removed, the average price level of the sector had risen sufficiently to allow costs to be covered. On the other hand, in a typical public enterprise, such as the electrical industry, price controls were still, to a certain extent, in effect in 1976 and, as a result, the industry was still operating at prices that forced it to make a loss, although the loss was considerably less than those made during the previous three years. In a few select industries, such as construction and leather, where there were relatively high degrees of competition and where controls were never strictly applied, there were no significant differences between zero profit and actual price levels until 1976, when the industries may have taken advantage of the general relaxation of controls to raise prices considerably above their zero-profit levels. Similar distortions in relative prices were found in most other sectors of the economy. The losses due to price controls peaked in 1974 and 1975 and were thought to have almost disappeared by the end of 1976.
In the analysis, the major sources of monetary expansion during 1963–76 were the government deficit, which included the government subsidies to public enterprises, and the estimated losses of private industries induced by price control. A time lag of approximately four quarters was found between the occurrence of losses and the subsequent creation of money. Additional sources of monetary expansion were changes in foreign reserves; the inflationary anticipations of the public were incorporated in a real balance equation. The magnitude of the estimated parameters indicated that, other things being equal, a one-point increase in the real government deficit as a percentage of the stock of real money balances caused, on the average, a 1.28 per cent point increase in the rate of growth in the monetary aggregate, an increase that may be interpreted as an estimate of the money multiplier. On the other hand, the estimation suggests that a one-point increase in the aggregate real loss of the private sector as a percentage of the stock of real money balances caused a 0.19 per cent point increase in the rate of growth in the money supply four quarters later. These percentages were significantly higher during the periods of higher rates of inflation.
Would inflation have increased at the same rate without such a large budget deficit and without price controls? The model was used to simulate three paths of inflation between 1967 and 1976 on the basis of three differing assumptions about the two policy variables: the government deficit and the losses in the public and private sectors induced by price control. This led to the following conclusions. The assumption of a balanced budget throughout the simulation period produced a simulated path of inflation that was significantly lower than the one generated by the assumption of the historical series of the two policy variables (see Chart 2). The average simulated quarterly rates of inflation declined from 5.30 to 2.23 per cent for the period 1967–72, from 34.77 to 5.86 per cent for 1973–76, and from 17.09 to 3.68 per cent for 1967–76. The assumption of no price distortions, and hence zero losses induced by controls for the public and private enterprises, also gave a simulated path of inflation markedly lower than the one generated by the assumption of the historical series of the policy variables, with average simulated quarterly rates of inflation being 1.13, 19.66, and 8.54 per cent for 1967–72, 1973–76, and 1967–76, respectively. According to this model, therefore, either a removal of price controls or a balancing of the government budget would have led to significant reductions in, although not a complete elimination of, the high rate of inflation that occurred in Argentina between 1967 and 1976.
Models on creation of credit, inflation, and real wages in Argentina—a technical note
The two studies were designed to test a series of hypotheses on sources of money creation. They were also used to simulate the rates of inflation and of changes in real wages that would have occurred under alternative assumptions about government policies, about deficit financing, and about the creation of credit to subsidize industries subject to price control. One study was based on a simple econometric model made up of (1) two stochastic functional relationships (that is, relationships containing elements of uncertainty), which explained the rate of change in the money supply and demand for real balances; (2) several definitional identities for connecting the nominal and real variables; and (3) an input-output system that generated an estimate of the losses in 17 of the 23 industries included in the Argentine input-output table for each quarter from 1963 to 1976.
The model, which was designed to capture the essence of the money creation-inflation spiral process described in the article, has the following features. The rate of increase in the monetary aggregate during the current quarter is largely determined by four variables: the current government deficit, the estimated aggregate loss of private industries during past quarters, current changes in external reserves (all in real terms), and the current and past rates of inflation. The rates of inflation, as determinants of the rate of monetary expansion, capture the degree of passivity of the changes in the money supply to the various pressures discussed earlier, while the use of estimated losses from previous quarters induced by price control is dictated by the institutional setting that allowed a time lag between the occurrence of losses and the actual creation of money.
Our method of estimating the business losses caused by price controls is relatively technical. Argentina has constructed input-output matrices representing the country’s production technology for the years 1963 and 1970. Input-output matrices for each of the years 1963–76 were constructed on the basis of these two matrices. The total values added, for our purposes given by the sectoral wage bills, were used to derive the Leontief, or zero-profit, prices for each sector in the input-output matrix for each of the 14 years in our study. These prices may be thought of as being those at which an average firm in each sector of the economy would have made exactly zero profit in the year in question. In a particular year, the difference between the computed price and the actual price is the distortion in the unit price of the sector in question. (In 1975–76 it is sometimes claimed that official and actual price indices differed, but in order to maintain consistency, we have chosen to use official price indices throughout the sample period.) If the actual price is lower than the computed zero-profit price, the distortion is positive, representing a loss to the sector. By multiplying this positive distortion by the total output of the sector, an estimate was obtained of the sector’s total loss caused by price controls. Totaling all sectors then gave an estimate of the aggregate loss due to price distortions for the entire economy. By subtracting transfer payments to public enterprises an estimate was made of the aggregate loss due to price controls that needed to be financed by domestic credit expansion. There is not, however, a one-to-one correspondence between these losses and net domestic credit, since not all enterprise losses actually received financing.
The second study examined the impact of inflation on real wage rates of several sectors of the industrial part of the Argentine economy. The design of the study was simple: a stochastic functional relationship was estimated to explain how the rate of change in nominal wage rates in each sector adjusted to the inflationary anticipation of the public and to simulate alternative paths of real wage rates on the basis of differing assumptions on government policies. One of the additional assumptions maintained throughout the study was that the public anticipated inflation only on the basis of its past experiences with inflation. This adaptive anticipation formula provides one possible reason why the real wage rate could fall even when the nominal wage rate fully adjusts to the anticipated rate of inflation. The estimated equations suggest, among other things, that the inflationary anticipation of the public was a dominant determinant of the changes in nominal wage rates in Argentina during 1963–76.
Chart 2Simulated quarterly rates of inflation based on different assumptions on policy
Inflation and real wages
It seems, then, that price controls probably did aggravate inflation in Argentina during the period of the study. Equally paradoxically, they also seemed to have helped to defeat another important objective that they aimed at—the maintenance of high real wages.
Real wages declined during the periods of high inflation, probably because nominal wages adjusted to inflationary anticipation rather than actual inflation, and because the public’s ability to forecast inflation accurately declined as inflation accelerated. The analysis also showed that an assumption of either a balanced government budget or zero losses from price controls in the public and private sectors would have raised the level of real wages significantly during the period. Thus, the policy of attempting to raise real wages by distorting the wage-price structure may have had significant counterproductive effects.
It is always tempting to overgeneralize the applicability of the results of a study such as this. Although price controls did seem to lead to inflation and an eventual lowering of real wages in Argentina between 1963 and 1976, it does not follow that such a sequence would necessarily occur elsewhere or even in Argentina at other times. The Argentine inflation contained critical and unique institutional elements, particularly the passivity of Central Bank credit. However, for Argentina, two important conclusions emerge from the study. First, and perhaps most important, price controls probably had an adverse impact upon the rate of inflation and the level of real wages, the two variables they were intended to improve. Second, the Central Government deficit was probably the major cause of high rates of inflation during the period in question.
Bijan B.Aghevli andMohsin S.Khan,“Government Deficits and the Inflationary Process in Developing Countries,”International Monetary Fund, Staff Papers, Vol. 25 (Washington, D.C., September1978).
PhillipCagan,“The Monetary Dynamics of Hyperinflation,” inStudies in the Quantity Theory of Money, edited byM.Friedman (Chicago: University of Chicago Press, 1956).
Ke-youngChu, andAndrewFeltenstein,“Relative Price Distortions and Inflation: The Case of Argentina, 1963–76,”International Monetary Fund, Staff Papers, Vol. 25 (Washington, D.C., September1978).
CarlosF. Diaz-Alejandro,Essays on the Economic History of the Argentine Republic (New Haven: Yale University Press, 1970).
Dean S.Dutton,“A Model of Self-Generating Inflation: The Argentine Case,”journal of Money, Credit and Banking, Vol. 3 (May1971).
Richard D.Mallon, andJuan V.Sourrouille,Economic Policymaking in a Conflict Society: the Argentine Case (Cambridge: Harvard University Press, 1975).